Predicting Upcoming Board Games

Predictive Models for BoardGameGeek Ratings

Pipeline

targets::tar_visnetwork(targets_only =T)

Warning message:
<anonymous>: ... may be used in an incorrect context: 
     class_metrics(group_by(add_pred_class(preds, threshold = .x), 
         threshold = .x), truth = {
         {
             outcome
         }
     }, estimate = .pred_class, event_level = "second", ...)
 

Models

averageweight_fit = 
    vetiver_pin_read(
        model_board,
        "bgg_averageweight_"
    )

average_fit = 
    vetiver_pin_read(
        model_board,
        "bgg_average_"
    )

usersrated_fit = 
    vetiver_pin_read(
        model_board,
        "bgg_usersrated_"
    )

hurdle_fit = 
        vetiver_pin_read(
        model_board,
        "bgg_hurdle_"
    )

Assessment

valid_predictions |>
    pivot_outcomes() |>
    left_join(
        games |>
            bggUtils:::unnest_outcomes() |>
            select(game_id, usersrated),
        by = join_by(game_id)
    ) |>
    plot_predictions(alpha = usersrated)+
    theme(legend.title = element_text())

targets_tracking_details(metrics = valid_metrics,
                         details = details) |>
    select(model, minratings, outcome, any_of(c("rmse", "mae", "mape", "rsq", "ccc"))) |>
    filter(minratings == 25) |>
    select(minratings, everything()) |>
    gt::gt() |>
    gt::tab_options(quarto.disable_processing = T) |>
    gtExtras::gt_theme_espn()

minratings	model	outcome	rmse	mae	mape	rsq	ccc
25	lightgbm	average	0.692	0.509	7.582	0.282	0.463
25	lightgbm	averageweight	0.457	0.347	19.210	0.665	0.804
25	lightgbm+lightgbm	bayesaverage	0.293	0.170	2.823	0.432	0.646
25	lightgbm	usersrated	1565.092	461.642	184.381	0.229	0.476

Tuning

Average Weight

Average Rating

Users Rated

Features

average_plot = 
    average_fit |> 
    extract_vetiver_features() |>
    plot_model_features()+
    labs(title = 'Average Rating')

averageweight_plot = 
    averageweight_fit |> 
    extract_vetiver_features() |>
    plot_model_features()+
    labs(title = 'Average Weight')

usersrated_plot = 
    usersrated_fit |> 
    extract_vetiver_features() |>
    plot_model_features()+
    labs(title = 'Users Rated')

Average Weight

Average

Users Rated

Predictions

# predict games
predictions = 
    upcoming_games |>
    impute_averageweight(
        model = averageweight_fit
    ) |>
    predict_hurdle(
        model = hurdle_fit,
        threshold = hurdle_threshold
    ) |>
    predict_bayesaverage(
        average_model = average_fit,
        usersrated_model = usersrated_fit
    )

Predictions for games expected to achieve at least 25 ratings.

# table
predictions |>
    filter(yearpublished >= 2024) |>
    filter(.pred_hurdle_class == 'yes') |>
    select(-starts_with(".pred_hurdle")) |>
    # this goddamn bah humbug game
    filter(game_id != 388225) |>
    predictions_dt(games = games) |>
    add_colors()

Probabilities for games on whether they will achieve 25 ratings.

predictions |>
    filter(yearpublished >= 2024) |>
    filter(.pred_hurdle_class == 'yes') |>
    arrange(desc(.pred_hurdle_yes)) |>
    filter(!is.na(thumbnail)) |>
    mutate(name = make_hyperlink(make_bgg_link(game_id), 
                                     mytext = paste(name, paste0("(",yearpublished, ")")))) |>
    mutate(Image = make_image_link(thumbnail),
            Game = name,
            Description = stringr::str_trunc(description, width = 150),
            `Pr(Hurdle)` = round(.pred_hurdle_yes, 3),
            Hurdle = hurdle,
            `Ratings` = usersrated,
            .keep = 'none') |>
    DT::datatable(escape=F,
                      rownames = F,
                      extensions = c('Responsive'),
                      class = list(stripe =F),
                      filter = list(position = 'top'),
                      options = list(pageLength = 15,
                                     initComplete = htmlwidgets::JS(
                                         "function(settings, json) {",
                                         paste0("$(this.api().table().container()).css({'font-size': '", '10pt', "'});"),
                                         "}"),
                                     scrollX=F,
                                     columnDefs = list(
                                         list(className = 'dt-center',
                                              visible=T,
                                              targets = c("Image", "Pr(Hurdle)", "Hurdle", "Ratings")
                                         )
                                     )
                      )
        )